[Dextrose] WiFi vs Suspend

[forster at ozonline.com.au]

Bernie

Sorry, on closer inspection its not locked up, it just looks like its locked up because it is doing something else for a long time.

Can be replicated 100%. Shut the lid for a while, 5 minutes is not long enough 1.5 hours is. Resume from sleep. The Wifi shows it is connected at 100% signal (signal is not 100%). There is no internet access. There is no response to Disconnect. Previously I had given up and restarted but if you wait 3 minutes or so it disconnects and then works normally.

Tony

> [cc += olpc-devel]
> 
> El Wed, 25-08-2010 a las 11:05 +1000, forster at ozonline.com.au escribió:
> > OS373pyg
> > Wifi is locked up after resume from sleep, must restart.
> > I presume this bug is being tracked at dev.laptop.org/ in one of the 4 tickets below
> 
> >   http://dev.laptop.org/ticket/10232  WiFi dies on suspended XO-1,os300
> >   http://dev.laptop.org/ticket/10092  Networking broken over suspend/resume on os13 for XO-1
> >   http://dev.laptop.org/ticket/9960   wake-on-WLAN doesn't always work\ (duplicate)
> >   http://dev.laptop.org/ticket/9967   ibertas suspend fails on XO-1 (fixed)
> 
> There are a number of unsolved bugs in the libertas kernel driver or in
> its fantastically proprietary firmare. If you turned on automatic power
> management on an XO-1, you might be seeing this other one:
> 
>  http://dev.laptop.org/ticket/10195
> 
> 
> In short, very rarely the libertas usb8xxx disappears from the USB bus
> when it receives certain commands from the driver. Suspend/resume are
> known to trigger quite frequently and the mesh also seems to cause it
> once or twice per day in a classroom of 30 students.
> 
> Because time for debugging was running out, in the latest beta builds I
> had to disable both mesh support and automatic power management in the
> attempt to get rid of this bug. After one week of testing, the problem
> was not reported any more.
> 
> If you spot this bug again, could you please:
> 
>  * check whether eth0 is still visible with "ifconfig -a"
> 
>  * check whether the Marvell 8xxx is still visible with "lsusb"
> 
>  * "dmesg >dmesg.out" and attach it to the bug. Dextrose enables
>    libertas debug in /etc/rc.local to help diagnose this bug.
> 
> 
> This never ending saga brings me back to an exasperated email that David
> Woodhouse wrote about 3 years ago, after spending many days in
> unfruitful debugging, in which he warned that access to the source code
> of the firmware was essential in order to nail down all the obscure
> Libertas bugs.
> 
> -- 
>    // Bernie Innocenti - http://codewiz.org/
>  \X/  Sugar Labs       - http://sugarlabs.org/
> 
> 
> _____________________________________________________
> This mail has been virus scanned by Australia On Line
> see http://www.australiaonline.net.au/mailscanning